[RightOnTheLeftCoast]

I used to teach TA in statistics as a grad student. The average is a horrible statistic. It is easily pulled by anomalous readings.

Example: What is the average of the following set: 1, 1, 1, 1, 10, 1?

The answer is easy enough to calculate: 2.5.

But is "2.5" representative of the set? What would be a typical value for the set? Is there any member of the set which is atypical and deserves another look?

Politicians love the average because it is such a fine tool for demagoguery. In the case of "average" temperature calculations, some readings come from peri-urban areas which have become increasingly developed in the past decades. Development means less foliage and brooks and more pavement. Significant increases in measured temperatures result, and these will pull any average calculated in a set which includes them.

[Raycpa]

Politicians love the average because it is such a fine tool for demagoguery.

I almost lost it during a presentation by someone from a representative of Habitat for Humanity. They were trying to make the lack of housing seem critical by using the following statistic:

The mean price for a house is X and almost 60% of the families in our state cannot afford it.

[Maneesh]

"The average is a horrible statistic. It is easily pulled by anomalous readings."

Excellent point. The distribution of the data us what determines the measure of central tendency that we should use and average can only be used when the data is normally distributed. The heat island effect in big cities can easily skew the data and make it non-normal, making averages essentially meaningless. The even bigger statistical flaw is that the models have omitted variation is water vapor levels which make up 95% of greenhouse gases. CO2 is less than 1% of greenhouse gases and man made CO2 is 9% of that 1%. Statistically if we considered all the potential factors for GW, there is no way that man made CO2 would ever be significant. The challenge is that our public education system produces people who can't do basic math in their head so to explain statistical variation to them...... This is truly an issue ripe for demagoguery and to liberals who think emotionally to begin with this is tailor made for obfuscation.

[bvw]

Thanks for the common-sense educational post.

Of course there is a number which is the average global temperature at any given time. But it is (1) MEANINGLESS and (2) UNKNOWABLE. Meaningless for any significant scientific study , meaningless for and economic, social or political decision. Unknowable to any degree of useful accuracy.

[ChessExpert]

The average American has one breast and one testicle. :)

[BIGLOOK]

Very interesting. Add to the factors of the set, the setting. Most recorded temperature readings are taken at airports or urban locales. Ma and Pa Kettle's thermometer readings out in the boondocks go unrecorded.

[mc5cents]

I used to teach TA in statistics as a grad student.

Ah, then you must be familuar with this quote by Mark Twain:

There are three kinds of lies: lies, damned lies, and statistics.

You can pretty much "prove" anything with satistics.

Don't make it true however.